|Publication number||US7844458 B2|
|Application number||US 11/554,086|
|Publication date||Nov 30, 2010|
|Priority date||Nov 2, 2005|
|Also published as||US20070100636|
|Publication number||11554086, 554086, US 7844458 B2, US 7844458B2, US-B2-7844458, US7844458 B2, US7844458B2|
|Inventors||Makoto Hirota, Hiroki Yamamoto|
|Original Assignee||Canon Kabushiki Kaisha|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (27), Referenced by (11), Classifications (7), Legal Events (2)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The present invention relates to a speech recognition apparatus and, more particularly, to a speech recognition apparatus using a multimodal user interface that is a combination of a graphical user interface (GUI) and a speech user interface (UI).
2. Description of the Related Art
Recent development of the speech recognition technology and improvement of hardware performance of speech recognition devices are enabling speech input in various computer-controlled devices such as car navigation systems, portable phones, and FAX apparatuses except personal computers and workstations.
Speech input generally provides the following merits.
(1) It allows a user to input without seeing the screen or using hands.
(2) It allows direct setting of items that are not displayed on the screen.
(3) It allows a user to set a plurality of items by one utterance.
Assume that a user wants to do copy setting in a copy machine to print a document on A4 sheets in quintuplicate. A normal GUI or UI based on key input makes the user to execute a plurality of steps, i.e., input the number of copies by using the ten-key pad, press the paper size button on the screen, and press the “A4” key on the screen.
With speech input, the user can set the paper size by only uttering “A4”, and this eliminates the user's effort required to display the paper size setting window as described in merit (2).
The user can also set the paper size and the number of copies at once by uttering, e.g., “A4, five copies” as described in merit (3).
There are also proposed techniques of increasing the operation efficiency by using multimodal input that combines GUI and speech instead of simply using GUI input or speech input (e.g., Japanese Patent Registration No. 2993872 and Japanese Patent Laid-Open No. 6-282569).
Although speech input has various merits, it also has a demerit of “misrecognition”. For example, even when the user utters “A4”, the speech recognition apparatus may misrecognize it as “A3”.
Even if the user utters “A4, five copies”, it may be misrecognized as “A4 to B5”. In this case, although the user wants to set two items, i.e., the paper size and the number of copies at once, the apparatus misrecognizes it as setting of one item “scaling factor”. The misrecognition of the item itself greatly confuses the user, who must then make much effort to correct the error.
The conventional GUI operation often uses hierarchical steps, i.e., makes the user select a setting item by a key and then set the detailed value of the setting item. This operation method can avoid misrecognition of the setting item itself, unlike speech input. However, the need to execute the plurality of steps increases the load on the user, as described above.
Even in multimodal input combining a GUI and speech input, a method using natural language analysis for speech input by a natural language, like Japanese Patent Registration No. 2993872, is susceptible to somewhat low accuracy of natural language analysis.
In view of the above problems in the conventional art, the present invention has an object to provide a speech recognition apparatus that enables efficient multimodal input by suppressing misrecognition in setting a plurality of items by one utterance.
According to one aspect of the present invention, a speech recognition apparatus for allowing setting by speech is provided. An input unit inputs a setting instruction by speech. A speech interpretation unit recognizes and interprets the contents of the setting instruction by speech to generate first structured data containing candidates of the interpretation result. An instruction input detecting unit detects a setting instruction input by a user. An instruction input interpretation unit interprets the contents of the setting instruction input to generate second structured data. A selection unit selects one of the interpretation candidates contained in the first structured data based on the second structured data.
Further features of the present invention will become apparent from the following description of exemplary embodiments with reference to the attached drawings.
Preferred embodiments of the present invention will be described in detail in accordance with the accompanying drawings. The present invention is not limited by the disclosure of the embodiments and all combinations of the features described in the embodiments are not always indispensable to solving means of the present invention.
In this embodiment, an example will be described in which setting of copy processing in a copy machine is done by speech input and touch panel input.
When the user taps the “number of copies” button, a ten-key input panel to set the number of copies appears, as shown in
As a characteristic feature of this embodiment, a plurality of items are efficiently input by using not only tap input on the LCD 204 but also speech input. An example will be described in which the user utters “A4, three copies” simultaneously while tapping the “paper size” button and “number of copies” button successively on the LCD 204 that displays windows shown in
The speech input unit 101 checks whether the user has input speech (S301). Speech input is generally detected on the basis of, e.g., the power of input speech. This is a known technique, and a detailed description thereof will be omitted. If YES in step S301, the speech recognition and interpretation unit 103 recognizes the input speech information (S302) and interprets the speech (S303). Speech recognition is a process of converting input speech into a character string. Speech interpretation is a process of converting the recognition result into predetermined structured data. When the user utters, e.g., “A4, three copies”, the speech is recognized, and “A4, three copies” is output by speech recognition processing (S302). This result is then converted into structured data containing a pair of an item name and an item value,
The process of recognizing and interpreting input speech and outputting it as structured data can be implemented by using an existing technology. The speech recognition and interpretation processing can use a grammatical rule described in, e.g., the speech recognition grammar specification SRGS or speech interpretation rule specification SISR of W3C (=World Wide Web Consortium).
Speech recognition processing can generally output top N results in descending order of likelihood (recognition score), instead of outputting only one result. Hence, N structured data of interpretation results can also be output.
The speech interpretation result is pushed to the speech interpretation stack provided in, e.g., the RAM 202 (step S304).
In the example shown in
Next, the tap input is interpreted (S403). This interpretation indicates output of data expressing that “setting item is XXX” by using the same structured data format as in the above-described speech interpretation. When the user taps the “paper size” button, data 110 shown in
The tap interpretation result is pushed to the tap interpretation stack provided, e.g., the RAM 202 (step S404).
If the user taps the “paper size” button and “number of copies” button successively, the tap interpretation stack holds data 120 next to the data 110, as shown in
Upon the processing described above with reference to the flowcharts in
If YES in step S502, the interpretation selection unit 104 selects, from the N interpretation candidates contained in the speech interpretation data, one interpretation candidate that most matches the tap interpretation data (S503). The matching degree is determined on the basis of the degree of matching between items. For example, the chronological order of the items contained in the tap interpretation data 110 and 120 in
On the other hand, if NO in step S502, the interpretation selection unit 104 determines that there is no tap interpretation data corresponding to the speech interpretation data. In this case, the interpretation selection unit 104 selects the top-ranked speech interpretation data candidate (SS04).
When one of the N speech interpretation data candidates is selected by the processing in step S503 or S504, the interpretation selection unit 104 clears both the speech and tap interpretation stacks (S505).
The GUI (display unit 105) displays the result of the above-described processing.
As described above, this embodiment allows to appropriately compensate speech recognition errors and obtain correct interpretation by integrating tap input with speech input. That is, in a UI to set a plurality of items, any risk of inputting an undesired value to an undesired item by a recognition error is avoidable while taking the advantage of speech input, i.e., inputting at least one item by one utterance.
If tap interpretation data corresponding to speech interpretation data is not present, the interpretation selection unit 104 of the first embodiment selects the top-ranked speech interpretation data candidate in step S504 in
The second embodiment of the present invention will be described below in detail. In the second embodiment, setting of copy processing in a copy machine is done by speech input and touch panel input, as in the above-described first embodiment. Especially as a feature, in searching for a speech recognition grammar on the basis of the feature parameter of input speech, candidates are narrowed down by tap input.
The hardware configuration of the copy machine according to the second embodiment is the same as in
An example will be described in which the user utters “A4, three copies” simultaneously while tapping the “paper size” button and “number of copies” button successively on an LCD 204 that displays GUIs shown in
The speech recognition grammar storage unit 1304 has a description to accept the pattern of simultaneous utterance of two settings for copy processing. Examples of the pattern of simultaneous utterance of two settings are
“A4, three copies”
“five copies, single-side to double-side”
“B5, double-side to double-side”
“at 141%, ten copies”
. . .
That is, the pattern inputs two arbitrary values of the four setting items “number of copies”, “paper size”, “double-side print”, and “scaling factor”.
First the speech input unit 1301 checks whether the user has input speech (S1401). If YES in step S1401, the feature extraction unit 1303 converts the input speech information into a feature parameter string (S1402). A known technique in the speech recognition processing field can implement the conversion to a feature parameter string. Feature parameters such as 01, 02, 03, . . . , On are output in chronological order.
The search unit 1305 selects, from the feature parameter string, the most appropriate phoneme sequence pattern that matches the speech recognition grammar stored in advance in the speech recognition grammar storage unit 1304 and outputs it as a recognition result (S1403). For example, assume that the speech recognition grammar is described as a pattern to accept speech input of the number of copies such as “one copy”, “two copies”, “three copies”, . . . In correspondence with this, phoneme sequence data as shown in
It is possible to create speech sequences expressing utterance patterns of two setting item values, as shown in
The user makes an utterance while tapping the “paper size” button and “number of copies” button successively. Hence, the tap input unit 1302 inputs tap information to the search unit 1305 when it is searching for one of the paths in
If the search unit 1305 obtains information representing that the user has tapped the “paper size” button at a given time during path search, the paths passing through a phoneme sequence 1602 of “paper size 1” in
Similarly, when the search unit 1305 obtains information representing that the second tap corresponds to the “number of copies” button, the paths passing through the phoneme sequence 1605 of “number of copies 2” include the correct path. Hence, the search unit 1305 excludes the paths passing through phoneme sequences 1606, 1607, and 1608 of “paper size 2”, “double-side print 2”, and “scaling factor 2” from the candidates for further search.
The interpretation unit 1306 converts the path found by the search unit 1305, i.e., the speech recognition result into structured data (S1404).
As described above, according to the second embodiment, upon receiving tap information during phoneme sequence pattern search processing, search candidates are narrowed down on the basis of the tap information. That is, the path closest to the feature parameter string is selected from paths matching the tap information. This allows to avoid any setting item input error due to a speech recognition error, as in the above-described first embodiment.
In the above-described example of the second embodiment, a speech recognition grammar to accept the pattern of simultaneous utterance of two settings for copy processing is prepared. The user simultaneously utters two settings and also taps the buttons of the two items.
Instead, the speech recognition grammar may have a description to accept a pattern to input an arbitrary number of settings by one utterance such that the user can simultaneously utter an arbitrary number of settings and also execute tap input an arbitrary number of times. In this case, the search unit selects search candidates matching the first tap information from phoneme sequences connected to a first node H in
In the above-described first and second embodiments, when the user taps one of setting item keys such as “number of copies”, “paper size”, “double-side print”, and “scaling factor”, a set panel of the item appears on the right side. For example, when the user utters “A4, three copies” simultaneously while tapping the “paper size” button and “number of copies” button successively, the “paper size” set panel and “number of copies” set panel appear on the screen successively.
However, if the user intends input by speech, such panel switching on the screen is meaningless and unnecessarily confuses the user. To solve this problem, a speech input ON/OFF button is provided. The speech input button in the ON state inhibits panel display switching or panel display even upon tapping a setting item key. ON/OFF control of speech input need not always be done explicitly by the key. For example, speech input may be set ON when the speech input unit detects input speech.
The above-described embodiment use speech input and tap input. Any user instruction input means other than speech input is applicable in place of tap input using the touch panel. For example, an operation such as mouse click on the display screen, pressing a predetermined key on the operation unit, or pressing a predetermined key provided on the keyboard is usable together with speech input in place of the tap input described in the embodiments.
Note that the present invention can be applied to an apparatus comprising a single device or to system constituted by a plurality of devices.
Furthermore, the invention can be implemented by supplying a software program, which implements the functions of the foregoing embodiments, directly or indirectly to a system or apparatus, reading the supplied program code with a computer of the system or apparatus, and then executing the program code. In this case, so long as the system or apparatus has the functions of the program, the mode of implementation need not rely upon a program.
Accordingly, since the functions of the present invention are implemented by computer, the program code installed in the computer also implements the present invention. In other words, the claims of the present invention also cover a computer program for the purpose of implementing the functions of the present invention,
In this case, so long as the system or apparatus has the functions of the program, the program may be executed in any form, such as an object code, a program executed by an interpreter, or scrip data supplied to an operating system.
Example of storage media that can be used for supplying the program are a floppy disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a CD-RW, a magnetic tape, a non-volatile type memory card, a ROM, and a DVD (DVD-ROM and a DVD-R).
As for the method of supplying the program, a client computer can be connected to a website on the Internet using a browser of the client computer, and the computer program of the present invention or an automatically-installable compressed file of the program can be downloaded to a recording medium such as a hard disk. Further, the program of the present invention can be supplied by dividing the program code constituting the program into a plurality of files and downloading the files from different websites. In other words, a WWW (World Wide Web) server that downloads, to multiple users, the program files that implement the functions of the present invention by computer is also covered by the claims of the present invention.
It is also possible to encrypt and store the program of the present invention on a storage medium such as a CD-ROM, distribute the storage medium to users, allow users who meet certain requirements to download decryption key information from a website via the Internet, and allow these users to decrypt the encrypted program by using the key information, whereby the program is installed in the user computer.
Besides the cases where the aforementioned functions according to the embodiments are implemented by executing the read program by computer, an operating system or the like running on the computer may perform all or a part of the actual processing so that the functions of the foregoing embodiments can be implemented by this processing.
Furthermore, after the program read from the storage medium is written to a function expansion board inserted into the computer or to a memory provided in a function expansion unit connected to the computer, a CPU or the like mounted on the function expansion board or function expansion unit performs all or a part of the actual processing so that the functions of the foregoing embodiments can be implemented by this processing.
While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.
This application claims the benefit of Japanese Patent Application No. 2005-320009, filed Nov. 02, 2005, which is hereby incorporated by reference herein in its entirety.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5748974 *||Dec 13, 1994||May 5, 1998||International Business Machines Corporation||Multimodal natural language interface for cross-application tasks|
|US6253184 *||Dec 14, 1998||Jun 26, 2001||Jon Ruppert||Interactive voice controlled copier apparatus|
|US6587820 *||Nov 16, 2001||Jul 1, 2003||Canon Kabushiki Kaisha||Information processing apparatus and method, a computer readable medium storing a control program for making a computer implemented information process, and a control program for selecting a specific grammar corresponding to an active input field or for controlling selection of a grammar or comprising a code of a selection step of selecting a specific grammar|
|US6633844 *||Dec 2, 1999||Oct 14, 2003||International Business Machines Corporation||Late integration in audio-visual continuous speech recognition|
|US6654720 *||May 9, 2000||Nov 25, 2003||International Business Machines Corporation||Method and system for voice control enabling device in a service discovery network|
|US6747678 *||Jun 15, 2000||Jun 8, 2004||Yamaha Corporation||Audio system, its control method and storage medium|
|US6823308 *||Feb 16, 2001||Nov 23, 2004||Canon Kabushiki Kaisha||Speech recognition accuracy in a multimodal input system|
|US7069215 *||Jul 12, 2001||Jun 27, 2006||At&T Corp.||Systems and methods for extracting meaning from multimodal inputs using finite-state devices|
|US7136909 *||Dec 28, 2001||Nov 14, 2006||Motorola, Inc.||Multimodal communication method and apparatus with multimodal profile|
|US7330868 *||Oct 8, 2003||Feb 12, 2008||Canon Kabushiki Kaisha||Data input apparatus and method|
|US7362465 *||Jul 1, 2002||Apr 22, 2008||Canon Kabushiki Kaisha||Information processing apparatus and control method therefor|
|US7363224 *||Dec 30, 2003||Apr 22, 2008||Microsoft Corporation||Method for entering text|
|US7383189 *||Apr 7, 2004||Jun 3, 2008||Nokia Corporation||Method and device for providing speech-enabled input in an electronic device having a user interface|
|US7421655 *||Jul 8, 2004||Sep 2, 2008||Canon Kabushiki Kaisha||Presenting information indicating input modalities|
|US7584427 *||Mar 19, 2001||Sep 1, 2009||Fuji Xerox Co., Ltd.||Operating method and device, and image processing apparatus using the same|
|US20030020760 *||Jul 5, 2002||Jan 30, 2003||Kazunori Takatsu||Method for setting a function and a setting item by selectively specifying a position in a tree-structured menu|
|US20030112277 *||Dec 14, 2001||Jun 19, 2003||Koninklijke Philips Electronics N.V.||Input of data using a combination of data input systems|
|US20040136023 *||Dec 12, 2003||Jul 15, 2004||Canon Kabushiki Kaisha||Method of controlling printing control apparatus|
|US20040172258 *||Dec 10, 2003||Sep 2, 2004||Dominach Richard F.||Techniques for disambiguating speech input using multimodal interfaces|
|US20050018658 *||Jul 8, 2004||Jan 27, 2005||Canon Kabushiki Kaisha||Information processing method and information processing apparatus|
|US20050139673 *||Dec 13, 2004||Jun 30, 2005||Canon Kabushiki Kaisha||Image forming apparatus, operation history storage method and control method, and storage medium|
|US20050283364 *||Jun 1, 2005||Dec 22, 2005||Michael Longe||Multimodal disambiguation of speech recognition|
|US20070005371 *||Jun 22, 2006||Jan 4, 2007||Canon Kabushiki Kaisha||Speech recognition method and speech recognition apparatus|
|US20080028327 *||Jul 25, 2007||Jan 31, 2008||Canon Kabushiki Kaisha||Information processing apparatus and user interface control method|
|US20100134677 *||Nov 24, 2009||Jun 3, 2010||Canon Kabushiki Kaisha||Image capturing apparatus, information processing method and storage medium|
|JP11993872A||Title not available|
|JPH06282569A||Title not available|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7983921 *||Jul 19, 2011||Canon Kabushiki Kaisha||Information processing apparatus for speech recognition with user guidance, method and program|
|US8782171 *||Jul 21, 2008||Jul 15, 2014||Voice Enabling Systems Technology Inc.||Voice-enabled web portal system|
|US9141335||Nov 21, 2012||Sep 22, 2015||Adobe Systems Incorporated||Natural language image tags|
|US20060293896 *||Jun 28, 2006||Dec 28, 2006||Kenichiro Nakagawa||User interface apparatus and method|
|US20080109223 *||Nov 6, 2007||May 8, 2008||Canon Kabushiki Kaisha||Information processing apparatus, method and program|
|US20080133238 *||Nov 30, 2007||Jun 5, 2008||Canon Kabushiki Kaisha||Information processing apparatus and information processing method|
|US20090024720 *||Jul 21, 2008||Jan 22, 2009||Fakhreddine Karray||Voice-enabled web portal system|
|US20100041479 *||Feb 18, 2010||Wei Hsu||Voice command game controlling apparatus and method of the same|
|US20120284031 *||Dec 9, 2010||Nov 8, 2012||Continental Automotive Gmbh||Method and device for operating technical equipment, in particular a motor vehicle|
|US20140078075 *||Nov 21, 2012||Mar 20, 2014||Adobe Systems Incorporated||Natural Language Image Editing|
|US20140082500 *||Nov 21, 2012||Mar 20, 2014||Adobe Systems Incorporated||Natural Language and User Interface Controls|
|U.S. Classification||704/252, 704/275, 715/727|
|International Classification||G10L15/22, G06F3/16|
|Oct 30, 2006||AS||Assignment|
Owner name: CANON KABUSHIKI KAISHA, JAPAN
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HIROTA, MAKOTO;YAMAMOTO, HIROKI;REEL/FRAME:018451/0629
Effective date: 20061026
|Apr 30, 2014||FPAY||Fee payment|
Year of fee payment: 4